Situation
A global cosmetics brand operated across retail, wholesale, and e-commerce channels, each with separate reporting systems and inconsistent sales definitions. Teams spent excessive time manually reconciling product-level sales across platforms including SAP HANA, Magento, and QlikView, leading to delays and decision friction at every monthly close.
Leadership lacked a single view of sales performance, making it difficult to analyze product trends, monitor channel dynamics, or track campaign effectiveness in real time. The fragmented environment meant that by the time numbers were reconciled and ready for review, the window for timely action had often closed.
Objective
We were asked to replace a manual, multi-system reporting process with a single source of truth that leadership could use to make decisions. The scope covered three channels (retail, wholesale, and e-commerce) running on different platforms, and required:
- Aligning sales definitions and SKU-level reporting logic consistently across all three systems.
- Consolidating multi-source sales data into a centralized, cloud-based warehouse with automated pipelines.
- Delivering interactive BI dashboards that surface product, region, and channel performance with drill-down capability.
- Integrating reconciliation logic to flag anomalies and resolve mismatches at the source, not after the fact.
Approach
We started by mapping the full data landscape: all three source systems, their field-level inconsistencies, and the manual steps teams were performing to bridge the gaps. That diagnostic shaped every design decision that followed.
Data standardizationWe standardized sales definitions and SKU-level reporting logic across SAP HANA, Magento, and QlikView. Aligning on a consistent definition of “gross sales,” “net sales,” and product hierarchy across three systems was the prerequisite that everything else depended on.
Automated data pipelinesWe built automated ingestion pipelines to pull from all three source systems into a centralized cloud warehouse on a near-real-time schedule. Reconciliation logic was embedded in the pipeline layer to flag mismatches and anomalies as data arrived, catching issues at ingestion rather than during end-of-month review.
BI dashboard suiteWe developed a suite of interactive dashboards covering quarterly sales performance and product mix trends, source-of-growth analysis by country, and SKU-level performance tracking across the top 20 product codes. Each dashboard supports region, channel, period, and category filters with full drill-down capability.
Figure 1Sales performance and product mix trends across FY18-FY20 by quarter and product type (Basic, Innovation, Limited Edition), with gross sales in $M and mix of business percentages, filterable by region, country, category, and channel.
Figure 2Source of growth by country (FY18-FY20 YTD): waterfall chart showing which markets drove year-over-year sales increases and which contributed to declines, giving leadership a clear view of geographic growth dynamics.
Figure 3Top 20 product code performance: gross profit, gross sales, mix of business, and YoY growth across FY18-FY20, with concentration risk assessed by comparing the top 20 against remaining SKUs across regions.
Impact
Manual reconciliation time dropped by over 60%, enabling faster monthly closes and freeing the analytics team from the error-prone, time-consuming work of bridging three disconnected systems by hand.
Business leaders gained a unified, accurate view of product-level sales across all channels and markets for the first time. The dashboards enabled real-time monitoring of promotions, product launches, and top-performing SKUs, giving marketing and commercial teams the visibility they needed to act on performance signals while they were still relevant.
The reconciliation logic embedded in the pipeline layer reduced downstream errors and gave data teams confidence in the numbers flowing into reports. The architecture was built to scale: adding new geographies or digital channels requires extending the pipeline rather than rebuilding the reporting layer from scratch.